In a new study, deepfake video clips of movie remakes that don't actually exist prompted participants to falsely remember the films—but simple text descriptions of the fake movies prompted similar false memory rates. Gillian Murphy of University College Cork, Ireland, and Lero, the Science Foundation Ireland Research Centre for Software, and colleagues presented these findings in the open-access journal PLOS ONE.
  • Deep fake videos can evoke false memories of the movies.
  • The study was conducted with the participation of 436 people.
  • Participants watched fake remake videos and text descriptions.
  • On average, 49% of participants believed that each fake remake was real.
  • The frequency of false memories from textual descriptions was just as high.
  • Deep forgery technology is no more effective than other memory corruption tools.
  • Participants expressed concern about the use of deep forgery technology.
  • Research shows that non-technical means can be effective in distorting memory.
      • WarmSoda@lemm.ee
        41·
        1 year ago

        I did? Are you trying to put false memories in my brain? That won’t work. I knew what I was watching.

  • Meowoem
    11·
    1 year ago

    Seems like a bit of a buzzword study, the interesting bit is participants expressing concern that ‘deep forgery technology’ was used because its totally meaningless. Why should anyone care which VFX tool was used? Unless they’ve been primed by media hype to think there’s some mysterious evil lurking…

    You tell someone about an old film and they think ‘yeah, that sounds like one of the million forgettable movies pumped out endlessly’ and maybe they even think ‘i might have seen that’ or ‘yeah that’s what that movie I saw twenty years ago with basically the same tropes was called’ - it’s meaningless, did generic action hero star in a movie called ‘true patriots’ or ‘verified ambition’ or ‘maximum attention’? I have literally no idea if they’re real films, or if I’ve seen them because they’re just totally generic - add a generic plot description or generic action scene and unless someone has a encyclopedic knowledge of movies they’re pretty much just guessing, hence the fifty-fifty result. Though they did it with remakes which seems even more likely to be ambiguous, Hollywood pumps out so many bad clones who is going to remember if any of the low grade cash grabs happened - will smith as neo is a stretch for sure but they made endless matrix junk no one saw, why wouldn’t they just randomly redo it with a black actor? He was in at least a dozen films with similar plots, some of which we remakes of older films - they appear to be asking university students so their pool probably trends to the early twenties with most participants not really very interested in the matrix or will smith, it’s not like they’re implanting memories that people have a strong opinion about or should know better.

    But no doubt this will get cited in fear spreading news articles as ‘deepfake AI implanted false memories and will doon us all!’ even though the actual research said it’s no better than any other method at tricking people into getting wrong answers on a movie memory quiz.